SN1 methylation produces many molecular lesions in DNA. About 7% of these are products of methylation at the O6 position of G (m6G), which are mutagenic, SCE-inducing, lethal, and very probably carcinogenic. While most normal cells contain MGMT (m6G-DNA methyltransferase) that repairs m6G by stoichiometric demethylation, our laboratory showed that about 20% of human tumor lines as well as 8 of 12 SV40 transformed human fibroblast lines fail to perform such repair. These repair-deficient cell lines display a number of phenotypic differences from repair- proficient lines including differential sensitivity to reproductive inactivation by chemotherapeutically used chloroethylating agents. Possibly germane to chemotherapy, biopsies from several brain tumors lacked completely any detectable capacity to repair m6G. Such findings have stimulated interest in understanding the relationship of improper repair of this lesion to both cancer cause and cure. While repair- proficient cells demethylate m6G, repair-deficient cells respond to m6G by incorporating nucleotides without removing the altered base. This repair-like response may be related to our finding that extracts of cultured human cell lines possess an activity that recognizes synthetic double-stranded oligodeoxynucleotides that contain one m6G:T base pair and incises them on either side of the poorly paired T. This response, similar biochemically to the repair of G:T mismatches in DNA, may initiate the abnormally high levels of sister-chromatid exchanges (SCEs), mutagenic, and lethal events observed in SN1-methylated repair-deficient cells. The study proposed here is designed to extend our understanding of the cellular response to m6G and to study the mechanism of G:T mismatch repair by extracts of human cells. Much of our proposed experimentation is based properties of A1235-MR4, which, by contrast to its parent, is resistant to MNNG-induced lethality and SCE, and to 6- thioguanine (6TG) - as well as BrdUrd-induced lethality. It shows a > 100x elevated spontaneous mutation frequency. To learn more about A1235- MR4, we will determine the bases in DNA at which spontaneous mutations occur. We will pursue our finding that, unlike extracts from other lines, extracts of MNNG-treated A1235-MR4 cells incorporate nucleotides into the one nucleotide gap produced by mismatch removal of T across from template m6G, and evaluate the possibility that the DNA polymerase beta of A12356-MR4 is altered. To evaluate mechanisms and mutants, we will study mismatch repair using substrates with base analogs. We propose to purify an inhibitor of m6G:T mismatch incision from HT29 cells that blocks incision of the m6G:T in preference to the G:T substrate. In addition we propose to determine whether we can select an MNNG-resistant mutant of adenovirus 5, which like A1235-MR4, may acquire such resistance by virtue of an altered DNA polymerase.